BMC Neuroscience BioMed Central

Research article Open Access The role of the in cell body enlargement, increased nuclear eccentricity and chromatolysis in axotomized spinal motor neurons David L McIlwain*1 and Victoria B Hoke2

Address: 1Department of Cell and Molecular Physiology, University of North Carolina School of Medicine, Chapel Hill, NC 27599, USA and 2Biomedical/Biotechnology Research Institute, North Carolina Central University, Durham, NC 27707, USA Email: David L McIlwain* - [email protected]; Victoria B Hoke - [email protected] * Corresponding author

Published: 17 March 2005 Received: 10 November 2004 Accepted: 17 March 2005 BMC Neuroscience 2005, 6:19 doi:10.1186/1471-2202-6-19 This article is available from: http://www.biomedcentral.com/1471-2202/6/19 © 2005 McIlwain and Hoke; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract Background: When spinal motor are injured, the , nucleus and cell body of the injured cell transiently increase in size, the nucleus becomes more eccentrically placed, and the organization of polyribosomes into Nissl bodies is temporarily disrupted. The mechanisms for these classical morphological responses to axotomy have not been satisfactorily explained. Results: In this study we address the role of the cell body cytoskeleton in these structural changes. We show that the cytoskeleton of uninjured lumbar cell bodies maintains nucleolar, nuclear and cell body size and nuclear position. When isolated, the relatively insoluble cell body cytoskeleton contains Nissl bodies and granules. After axotomy, protein labeling increases markedly and the cytoskeleton enlarges, increasing nucleolar, nuclear and cell body size, as well as nuclear eccentricity. Nearly all of the protein mass that accumulates in the cell body after axotomy appears to be added to the cytoskeleton. Conclusion: We conclude that axotomy causes the conjugate enlargement of the nucleolus, nucleus and cell body and increases nuclear eccentricity in spinal motor neurons by adding protein to the cytoskeleton. The change in nuclear position, we propose, occurs when cytoskeletal elements of the cannot enter the shortened axon and "dam up" between the nucleus and . As a consequence, we suggest that -free axonal cytoskeleton accumulates between the nucleus and axon, displaces Nissl body-containing cytoskeleton, and produces central chromatolysis in that region of the cell.

Background known and explored for over a century, these morpholog- Injury to the axon of spinal motor neurons produces ical responses to axotomy have not been adequately major structural changes in the affected cell body. These explained. changes include a transient enlargement of the nucleolus, nucleus and cell body, an increase in nuclear eccentricity It has been proposed that enlargement of the motor neu- and central chromatolysis, which is a centrifugal disap- ron cell body after axotomy is caused by the entry of water pearance of polyribosome-containing Nissl bodies. While into the cell bodies, possibly the result of an early increase

Page 1 of 13 (page number not for citation purposes) BMC Neuroscience 2005, 6:19 http://www.biomedcentral.com/1471-2202/6/19

in osmotically-active hydrolytic products of macromole- cules within the motor neurons [1]. On the other hand, the gradual increase in total protein and RNA observed in 6000 axotomized motor neurons cell bodies [2,3] could also 5000 Unextracted play a role in their enlargement after injury. That the Extracted ) nucleolus, nucleus and cell body enlarge together after 3 4000 axotomy suggests that their sizes may be coordinated, 3000 …. especially since a similar scaling is observed in normal, . 2000 non-injured motor neurons of increasing size [4] and in Nucleus (µm motor neurons exposed to excess growth hormone [5]. It 1000 is noteworthy that the nucleus, nucleolus and cell body do 0 not each enlarge after axotomy in all types of neurons 0 10000 20000 30000 40000 capable of axon regeneration, implying that an increase in Cell body volume (µm3) size may not be required for regrowth of the axon [6].

240 The repositioning of the nucleus in injured motor neu- Unextracted

) 200 rons was quantified by Barr and Hamilton [1], who, like 3 Extracted others [7,8], found that the direction of increased nuclear 160 eccentricity in motor neurons was usually away from the 120 axon hillock. Suggested mechanisms for increased nuclear eccentricity include a selective influx of water into the area 80

of the axon hillock [1] and interference with axonal trans- Nucleolar volume (µm 40 port, leading to the accumulation of axonal constituents 0 in the injured cell body [6]. 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 Nuclear volume (µm3) Central chromatolysis also develops progressively in the region between the nucleus and axon hillock after axot- omy [6]. The gradual disappearance of Nissl bodies from PreservationtionshipsFigure 1 in extracted of cell body, spinal nucle motorar and neurons nucleolar volume rela- the perinuclear region towards the periphery of the cell Preservation of cell body, nuclear and nucleolar vol- body and the concomitant loss of basophilic staining of ume relationships in extracted spinal motor neurons. RNA occur even as the radiolabeling and total content of Nuclear volume in individual cell bodies isolated from unfixed RNA increase [6]. Neither the molecular basis of chroma- lumbar spinal cords of normal adult frogs scales with cell tolysis nor the reason for its appearance and centrifugal body volume (top) and nucleolar volume (bottom) before spreading between the nucleus and axon hillock is and after extraction. Significant (p < 0.01) correlation coeffi- known. cients were found in first-order regression analyses for both unextracted (solid line) and extracted cells (dashed line). In this study, we first show that the size and shape of nor- mal, non-injured frog motor neuron cell bodies are main- tained by a cytoskeleton that can be isolated. We then provide evidence that axotomy increases the synthesis and total content of proteins in the cytoskeleton, which range of cell body areas in motor neurons isolated from appears to alter its structure and produce nucleolar, normal animals by the method of Sinicropi and McIlwain nuclear and cell body enlargement. We further show that [10] is slightly larger than that of their fixed, sectioned the cytoskeleton maintains nuclear position in the unin- counterparts [5]. As with area measurements on normal jured cell body and is altered by axotomy to increase motor neurons [5], nucleolar and nuclear volume increase nuclear eccentricity. Finally, we propose that both nuclear as cell body volume increases (Fig. 1). Since the plasma eccentricity and central chromatolysis result from an accu- membrane is damaged during the isolation of motor neu- mulation of axonal cytoskeleton that is impeded from ron cell bodies and is no longer semipermeable to most entering the truncated axon. osmolytes [11], osmotic forces do not appear to be chiefly responsible for maintaining cell body size in isolated Results motor neurons. Even when intracellular membranes are The cytoskeleton maintains nucleolar, nuclear and cell permeabilized with 1% Triton X-100 and over one-half of body size in normal motor neurons the cell body protein is removed from isolated motor neu- Isolated, unfixed motor neurons, like fixed, sectioned rons by successive extraction and high salt solutions, the motor neurons, vary widely in cell body area [5,9]. The mean size of the nucleolus, nucleus and cell body is little

Page 2 of 13 (page number not for citation purposes) BMC Neuroscience 2005, 6:19 http://www.biomedcentral.com/1471-2202/6/19

Table 1: The cytoskeleton maintains most of cell body, nuclear and nucleolar volume in normal motor neurons.

Unextracted Volume (µm3) Extracted Volume (µm3)E/Ua

Nucleolus 74.5 ± 11.0 86.8 ± 18.3 1.17 Nucleus 2,402 ± 423 1,968 ± 339** 0.82 Cell Body 19,387 ± 2,160 16,409 ± 1,662** 0.85

Mean volume ± S.D from 6 experiments; approximately 50 lumbar cell bodies from 3 frogs were analyzed in each experiment. Isolated cell bodies were extracted by the method of He et al. [12] to obtain . a E/U = extracted/unextracted **p < 0.01, Student's paired t-test, extracted vs. unextracted counterpart

affected (Table 1). This extraction procedure, which has been used to isolate the nuclear matrix and cytoplasmic skeleton of non-neuronal cells [12], reduces the cell body protein content in normal frog motor neurons from 4.35 ± 0.87 to 1.88 ± 0.29 ng/cell (a 56.8% decrease). In con- trast, cell body and nuclear volume decline only by 15 and 17%, respectively, and nucleolar volume does not change significantly (Table 1). The volumes of the nucleolus and nucleus in the extracted motor neurons continue to scale with cell body volume (Fig. 1) and the extracted cells have a relatively normal appearance by differential interference contrast microscopy (Fig. 2). These findings suggest that nucleolar, nuclear and cell body size in motor neurons are maintained by cytoskeletal structures and led us to exam- ine the ultrastructure of extracted motor neurons.

Resinless thin sections [12] were prepared from frog lum- bar spinal cord that had been extracted in situ by the same procedure used for isolated cell bodies. Isolated, extracted cell bodies were not used, because they produce friable thin sections with this embedment procedure. Transmis- sion electron microscopy of motor neurons in unstained thin sections from which the embedment medium – diethyleneglycol distearate (DGD) – had been removed reveals a three-dimensional network of filamentous struc- tures that remains after extraction of soluble proteins (Fig. 3). The nuclear skeleton of motor neurons is much less densely organized than the cytoplasmic skeleton, which is a highly cross-linked network of filamentous proteins with electron dense particles – possibly ribosomes [13] – interspersed throughout the lattice. The filamentous struts of the lattice are tapered, suggesting the possibility of arti- LightspinalFigure microscopicmotor 2 neurons appearance of extracted and unextracted factual deposition of some protein onto the lattice. These Light microscopic appearance of extracted and unex- images, together with the morphometric analyses tracted spinal motor neurons. Differential interference described above, indicate that the cytoskeleton is largely contrast images of lumbar motor neuron cell bodies isolated responsible for maintaining nucleolar, nuclear and cell from normal adult frogs. Panel a shows an unextracted cell body size in motor neurons. body. Panel b shows a different cell body isolated from nor- mal adult frogs and extracted by a modification of the proce- dure of He et al. [12]. Each bar = 20 µm. An interesting aspect of these data is the influence of the embedment medium in thin sections, which is illustrated

Page 3 of 13 (page number not for citation purposes) BMC Neuroscience 2005, 6:19 http://www.biomedcentral.com/1471-2202/6/19

motor neurons. To determine whether Nissl bodies are associated with the isolated cytoskeleton, we examined cell body cytoskeletons isolated from human motor neu- rons and stained with methylene blue. Confocal images of methylene blue-stained cytoskeletons from normal human motor neuron cell bodies showed abundant Nissl bodies in the cell body and proximal dendrites (Fig. 5). As with isolated frog motor neurons, the appearance of iso- lated human cell bodies is relatively unchanged after extraction.

Adult human motor neurons, unlike those of frog, also contain large amounts of lipofuscin, and cytoskeletons isolated from human motor neuron cell bodies retain their lipofuscin (Fig. 5). The possibility that lipofuscin granules are trapped within the interstices of the cytoskel- eton of the cell body and dendrites became apparent when isolated human cell bodies being examined by dif- ferential interference contrast microscopy (DIC) were exposed to SDS. When concentrations of up to 1% of the detergent were introduced at the edge of the coverslip, the cell bodies often began to swell and, one by one, individ- ual lipofuscin granules were released and floated away from the enlarged cell body. The granules had average ElectroncytoskeletonFigure 3 microscopic appearance of the spinal motor neuron diameters of about 2.5 m (range 1.8–3.0 m, n = 16) and Electron microscopic appearance of the spinal motor µ µ neuron cytoskeleton. Embedment-free electron micro- appeared to be insoluble in SDS. Even after depletion of graphs of two motor neurons within frog lumbar spinal cord most of the cell's lipofuscin, the swollen motor neuron extracted as described in Methods. The extracted spinal was sometimes still visible by DIC. Lipofuscin was cord, from which the motor neurons were not isolated, was released only from cell bodies that underwent swelling. fixed with 2.5% glutaraldehyde and embedded and sectioned This sequence of events led us to infer that lipofuscin in DGD. After removal of the embedment medium and granules may be caged within the cell body cytoskeleton repeated drying through the critical point, the sections were and can individually escape it as the lattice enlarges to placed on Formvar- coated grids without staining and visual- pore sizes that exceed the diameter of each granule. ized by transmission electron microscopy. Nu = nuclear skel- eton; nl = nuclear lamina; cy = cytoplasmic skeleton. Inset bar Although the swollen cell bodies eventually became invis- = 10 µm; at higher magnification, bar = 250 nm. ible by light microscopy, the cytoskeleton may not be completely soluble in SDS. Approximately one-third of the cell body protein (31.0 ± 15.1%, n = 5) was recovered in the pellet when spun at 13,000 × g for 5 min in the pres- ence of 1% SDS. This may explain our inability thus far to in Fig. 4. The three-dimensional lattice seen when the detect and identify any proteins in the isolated cytoskele- embedding medium is removed is completely invisible ton by SDS-polyacrylamide gel electrophoresis. when the embedding medium is present (Fig. 4, middle panel) and differs greatly from the two-dimensional Axotomy enlarges the cell body cytoskeleton image of stained cells embedded in plastic (Fig. 4, right Transection of motor axons in the frog lumbar ventral panel), which masks much of the underlying filamentous root causes progressive enlargement of the nucleolus, structure. The obscuring influence of the embedding resin nucleus and cell body of the injured motor neurons [4]. has been discussed by Penman [14]. Given the evidence above that the cytoskeleton largely maintains normal nucleolar, nuclear and cell body size, Nissl bodies and lipofuscin are associated with the isolated one would postulate that axotomy increases their size by cytoskeleton altering the cytoskeleton. This possibility was tested in Motor neurons isolated from the grass frog stain with morphometric analyses of cytoskeletons isolated from basophilic dyes before and after extraction, indicating the motor neuron cell bodies of frogs 16 days following ven- presence of ribosomes. However, they do not contain tral root transection, when injury-induced enlargement well-defined Nissl bodies, such as those seen in human was well underway. As in normal motor neurons, the

Page 4 of 13 (page number not for citation purposes) BMC Neuroscience 2005, 6:19 http://www.biomedcentral.com/1471-2202/6/19

EffectFigure of 4 embedment media on the ultrastructural appearance of the motor neuron cytoskeleton Effect of embedment media on the ultrastructural appearance of the motor neuron cytoskeleton. Three views of the motor neuron cell body cytoskeleton. Left: the unstained thin section from Fig. 3 of extracted spinal tissue from which DGD was removed; Middle: an unstained thin section of extracted spinal tissue from which DGD was not removed; Right: an Epon-embedded thin section of extracted spinal tissue, post-stained with uranyl acetate and lead citrate. Symbols as in Fig. 3. The arrows denote the extracted nuclear envelope. Each bar = 250 nm.

cytoskeleton of axotomized cells maintains most of nucle- olar, nuclear and cell body size (Table 2). Scaling of nucle- olar, nuclear and cell body sizes after injury was also evident in both unextracted and extracted cell bodies (Fig. 6). By postoperative day 16, mean nucleolar, nuclear and cell body volumes had increased 222%, 148% and 145% in the isolated cytoskeletons, respectively, versus 254%, 159% and 164% in unextracted cell bodies (Tables 1 &2). Dividing the injury-induced increases in volume in extracted cells by those in unextracted cells, one finds that the enlargement of the cytoskeleton accounted for 91% of the increase in nucleolar size, 67% of nuclear size, and 60% of cell body size.

The injury-induced expansion of the motor neuron cell body is not uniform in all dimensions. In both unex- tracted and extracted cell bodies, axotomy enlarged the x, y axis much more than the z-axis of the cell. For example, in 5 experiments on a total of 145 unextracted, 16-day axotomized cell bodies, the mean radius increased 26.6 ± 4.5% after injury (p < 0.01; unpaired Student's t-test), while the mean height (z-axis) of the cells increased by RetentionneuronsFigure 5 of Nissl bodies and lipfuscin by extracted motor only 7.3 ± 8.4% (p = n.s.). Similar differences were found Retention of Nissl bodies and lipfuscin by extracted in cytoskeletons isolated from 16-day axotomized cell motor neurons. Confocal image of an extracted motor bodies. neuron cell body isolated from human lumbar spinal cord and stained with methylene blue. The pink structures (small Axotomy adds protein selectively to the cell body arrows) are Nissl bodies and the blue and yellow structures cytoskeleton are lipofuscin granules (large arrow). Bar = 25 µm. Following ventral root transection, the total protein con- tent of the injured motor neuron cell body increases

Page 5 of 13 (page number not for citation purposes) BMC Neuroscience 2005, 6:19 http://www.biomedcentral.com/1471-2202/6/19

Table 2: The cytoskeleton maintains most of cell body, nuclear and nucleolar volume in axotomized motor neurons.

Unextracted Volume (µm3) Extracted Volume (µm3)E/Ua

Nucleolus 189.3 ± 37.9 191.7 ± 38.8 1.01 Nucleus 3,825 ± 864 2,916 ± 513* 0.76 Cell Body 31,717 ± 4,299 23,831 ± 3,381* 0.75

Mean volume ± S.D from 6 experiments on 16-day axotomized motor neurons. Approximately 50 lumbar cell bodies from 3 frogs were analyzed in each experiment. a E/U = extracted/unextracted *p < 0.05, Student's paired t-test, extracted vs. unextracted counterpart

steadily with increasing cell body volume (Fig. 7). When we compared the protein content of unextracted cell bod- 12000 ies axotomized 16 days earlier to that of their isolated cytoskeletons, we found that virtually all of the protein 10000 Unextracted ) 3 Extracted added to the cell body after axotomy was recovered in the 8000 isolated cytoskeleton (Table 3). These data imply that the

6000 addition of protein to the cell body cytoskeleton is responsible for its enlargement after axotomy. 4000 3 Nuclear volume (µm 2000 When normal motor neurons were labeled in situ with H- leucine and their cell bodies isolated and extracted, 0 slightly over one-half of the newly synthesized protein in 0 20000 40000 60000 80000 Cell body volume (µm3) normal motor neuron cell bodies was recovered in the iso- lated cytoskeleton (Table 4). Axotomy increased protein 800 labeling six-fold in both unextracted cell bodies and their Unextracted isolated cytoskeletons by postoperative day 16, and again ) 3 Extracted 600 about one-half of the total labeled protein was recovered in the isolated cytoskeleton (Table 4). Thus, in normal and axotomized cell bodies about one-half of the newly 400 synthesized protein is not found in the isolated cytoskele- ton. However, only new protein added to the cytoskeleton 200 after axotomy is associated with the increase in total pro- Nucleolar volume (µm tein content of the cell body (Tables 3 and 4). Newly 0 synthesized protein not recovered with the cytoskeleton, 0 3000 6000 9000 12000 15000 although elevated by injury, does not appear to contribute 3 Nuclear volume (µm ) measurably to the increase in total protein content of the cell body after axotomy, indicating that it is exported and/ PreservationtionshipsFigure 6 in axotomized, of cell body, extracted nuclear and motor nucleolar neurons volume rela- or degraded more rapidly than most protein in the iso- Preservation of cell body, nuclear and nucleolar vol- lated cytoskeleton. ume relationships in axotomized, extracted motor neurons. Motor neurons enlarged by axotomy also exhibit Axotomy increases nuclear eccentricity by altering the scaling of nucleolar, nuclear and cell body volume before and cytoskeleton after extraction. Nuclear volumes in cells isolated 16 days Based upon evidence that the cytoskeleton maintains after transection of frog lumbar ventral roots correlate with nuclear location in non-neuronal cells [15], we analyzed cell body volume (top) and nucleolar volume (bottom) over a nuclear position in cytoskeletons isolated from non- wide range of cell body size. Correlation coefficients from injured, extracted motor neuron cell bodies. Extracted frog first-order regression analyses of unextracted (solid lines) motor neurons showed the same degree of nuclear eccen- and extracted cells (dashed lines) were significant at p < 0.05. tricity found in unextracted cell bodies (Table 5). Approx- imately 10% of the normal cell bodies, whether extracted or unextracted, had nuclei that contacted the cell periph- ery (100% eccentricity). After ventral root transection, nuclear eccentricity in the isolated cytoskeleton increased

Page 6 of 13 (page number not for citation purposes) BMC Neuroscience 2005, 6:19 http://www.biomedcentral.com/1471-2202/6/19

Table 3: Selective addition of protein to the cell body cytoskeleton after axotomy.

Total Protein (ng/cell) Added Protein (ng/cell) Normal Axotomized

Unextracted 4.35 ± 0.87 6.51 ± 2.44* 2.16 Extracted Cells 1.88 ± 0.29†† 4.07 ± 0.52**† 2.19

The total cell body protein content of isolated cell bodies and their cytoskeletons is expressed as a mean ± S.D. (n = 4–9 experiments) for groups of cell bodies isolated from normal and 16-day axotomized frogs. * p < 0.05; ** p < 0.01: axotomized vs. normal, Student's unpaired t-test † p < 0.05; †† p < 0.01: extracted vs. unextracted, Students unpaired t-test

10

R2 = 0.9837 20 days 8

16 days

12 days 6

Protein (ng/cell body) 8days 0 days

4 15 20 25 30 35 40 45 Cell Body Volume (um3x10-3)

ConcurrentFigure 7 increases in motor neuron cell body volume and protein content after axotomy Concurrent increases in motor neuron cell body volume and protein content after axotomy. Injury-induced increases in mean cell body volume and protein content between 8 and 20 days after axotomy are closely correlated (p < 0.01) in third-order regression analyses. Error bars represent standard error of the mean.

Page 7 of 13 (page number not for citation purposes) BMC Neuroscience 2005, 6:19 http://www.biomedcentral.com/1471-2202/6/19

Table 4: Newly synthesized protein in isolated cell body cytoskeletons before and after axotomy.

3H-leucine-labeled protein (cpm/cell body) Normal Axotomized

Unextracted cells 36.0 ± 15.4 210.6 ± 25.7** Extracted Cells 19.8 ± 11.1† 114.8 ± 29.1**†† E/U × 100a 55.0% 54.5%

Mean cpm ± S.D. for cell bodies and their cytoskeletons isolated from labeled spinal cords of normal and 16-day axotomized frogs in 3 experiments. a E/U × 100 = percent new protein allocated to the cell body cytoskeleton ** p < 0.01: axotomized vs. normal, Student's unpaired t-test † p < 0.05; †† p < 0.01: extracted vs. unextracted, Students paired t-test

Table 5: Nuclear position is maintained by the cytoskeleton in normal and axotomized motor neurons.

Nuclear Eccentricity (%)a Unextracted Extracted E/Ub

Normal 54 ± 6 61 ± 4 1.13 Axotomized 84 ± 8** 86 ± 9** 1.02

a Nuclear eccentricity, expressed as a mean percent ± S.D, was determined by the method of Barr and Hamilton [1]. For each mean value shown, 98–139 cell bodies were analyzed (n = 5 experiments). At 0% eccentricity, the nuclear and cell body centers are superimposed; at 100%, the nuclear perimeter contacts the cell body perimeter. b E/U = extracted/unextracted **p < 0.01, Student's unpaired t-test, axotomized vs. normal

to the same degree as it did in unextracted cell bodies and 2) and the tapered appearance of its filaments (Fig. 3) (Table 5). Approximately 60% of the injured cell bodies probably reflect those changes. Nissl bodies might not be had nuclei that touched the cell periphery, whether or not associated with the cytoskeleton in vivo, although they were extracted. Thus, the cytoskeleton maintains cytoskeletons isolated in a similar way from non-neuronal nuclear position in normal motor neurons, as it does mammalian cells also contain polyribosomes [13,16,17], nuclear volume, and is altered by axotomy to produce an and in plant cells, labeled polyribosomes added prior to increase in nuclear eccentricity. the extraction procedure were not adsorbed by the cytoskeleton [18]. Discussion The cell body cytoskeleton in normal spinal motor neurons Role of the cytoskeleton in the conjugate enlargement of The ability to isolate individual neuronal cell bodies sim- the nucleolus, nucleus and cell body after axotomy plifies volumetric measurements of the entire cell body The data in this study form the basis for a new explanation and its cytoskeleton and permits one to quantify their for the events leading to the transient enlargement of the total protein content and radiolabeling. The isolated cell nucleolus, nucleus and cell body after axotomy. We find body cytoskeleton retains the shape and most of the size that axonal injury adds protein to the cytoskeleton, likely of the unextracted motor neuron. It maintains the size enlarging the cell body, as well as the nucleus and nucle- relationships among the nucleolus, nucleus and cell body olus, which are integral parts of its filamentous structure. over a wide range of cell body volumes and fixes the posi- The injury-induced accumulation of proteins in the iso- tion of the nucleus within the cell. lated cytoskeleton implies that their rate of synthesis exceeded their rate of loss from the cell body. This condi- The morphological characteristics of the isolated cytoskel- tion could result from an increase in their rate of synthe- eton are so typical of motor neurons in situ, that it is diffi- sis, a decrease in their rate of degradation or export from cult to conceive that the cytoskeletons we isolate do not the cell body or a combination of these factors. The injury- exist in the cells from which they are derived. However, induced increase we observed in labeled protein in the our experimental procedures do cause changes in the liv- isolated cytoskeleton does not distinguish among these ing cell bodies. The somewhat smaller volumes of the possibilities. However, the rate of total protein synthesis in nucleus and cell body in isolated cytoskeletons (Tables 1 frog motor neurons is probably increased after axotomy,

Page 8 of 13 (page number not for citation purposes) BMC Neuroscience 2005, 6:19 http://www.biomedcentral.com/1471-2202/6/19

since we have previously described multiphasic increases ing the high and middle molecular weight in total transcription during this postoperative period [4]. subunits are incapable of producing , but It is unclear how much of that increase is directed towards still retain their normal shape. Likewise, Larivière et al. proteins in the cytoskeletal and non-cytoskeletal fractions, [23] found that motor neurons in transgenic mice lacking but the fact that each of these two fractions receives about peripherin also have normal-appearing cell bodies. one-half of the total label both before and after axotomy Although compensations for losses of the individual pro- suggests that the rate of protein synthesis increases in both teins could conceivably preserve cell body structure, these fractions. We can also infer that the relationships between observations call into question whether neurofilaments or protein synthesis, degradation and export differ in the peripherin are essential for much of the shape of normal cytoskeletal and non-cytoskeletal fractions of the injured motor neuron cell bodies. The identification of the major motor neuron cell bodies. As noted, the increase in structural proteins in the cytoskeleton of motor neuron labeled protein in the isolated cytoskeleton is cell bodies will depend upon how amenable its proteins accompanied by an increase in protein mass, indicating are to current modes of protein compositional analysis. that the rate of addition of new protein to the cytoskeleton is not matched by an increase in the rate of protein loss Role of the cytoskeleton in increased nuclear eccentricity from it. In contrast, the injury-induced increase in labeled and central chromatolysis after axotomy protein that is not associated with the isolated cytoskele- We also show that changes in the cell body cytoskeleton ton produces no significant protein accumulation in that appear to be responsible for the increase in nuclear eccen- fraction, indicating equal rates of protein synthesis and tricity following axonal injury. An asymmetric addition of loss after axotomy. These observations illustrate once protein to the cytoplasmic skeleton cannot alone account again why it is unsafe to use labeled protein as an indica- for the change in nuclear position, because the fraction of tor of protein mass. axotomized cells with nuclei touching the periphery of the cell body (100% eccentricity) is about six times that found Are neurofilaments and peripherin major determinants of in normal cells. Thus, axotomy not only enlarges the cell body size and shape in motor neurons? cytoskeleton, but also causes dissolution of the lattice on An intriguing question raised by our study pertains to the one side of the cell. identity of the cytoskeletal constituents of the filamentous lattice in spinal motor neurons. On the one hand, there is Additional clues to the mechanism by which the nucleus evidence that cell body size in neurons can be changed by is relocated in the cell are found in the direction of nuclear manipulating its intermediate filament content. For displacement and the location of chromatolysis after axot- example, increased cell body size in transgenic rat motor omy. As mentioned, several investigators have reported neurons overexpressing the high molecular weight com- that the direction of nuclear displacement is away from ponent of neurofilaments (NF-H) can be reduced by con- the axon hillock [1,7,8]. We did not test this directionality currrently increasing the expression of the low molecular here, because the axon is routinely lost during cell body weight component (NF-L) [19]. Likewise, cell body length isolation. Many others have reported that the loss of Nissl in differentiated PC-12 cells changes when peripherin is bodies (i.e., chromatolysis) also appears first in the region inhibited by transfecting the cells with small interfering of the cell between the nucleus and the axon hillock [6]. If mRNAs [20]. Conversely, axotomy-induced cell body Nissl bodies are associated with the cell body cytoskeleton enlargement is accompanied by increases in the cell body in vivo, as they are in the isolated cytoskeleton, then one content of NF-L in frog motor neurons [10] and increased must ask why Nissl bodies are not present in the peripherin immmunoreactivity in rat spinal motor neu- cytoskeleton added between the nucleus and axon after rons [21]. Unlike neurofilament protein, peripherin axotomy. There is no reason to believe that axotomy mRNA also increases after axotomy [21], a finding con- causes Nissl bodies to be removed only from the cytoskel- sistent with the abovementioned possibility of an eton added to that particular region of the cell. A more increased rate of synthesis of cytoskeletal proteins in our plausible explanation is that when the axon, which does study. However, if neurofilaments or peripherin are con- not contain Nissl bodies [24], is drastically shortened by stituents of the isolated cytoskeleton, we predict that they axotomy, it cannot incorporate all of the axonal cytoskel- are cross-linked (Fig. 3) or otherwise modified so as to eton being synthesized by the injured cell body. Injury- reduce their solubility. induced impairment of axonal transport could exacerbate this problem. The excess of Nissl body-free axonal On the other hand, there is evidence that neither neurofil- cytoskeleton could then accumulate between the nucleus ament proteins nor peripherin are responsible for and axon hillock, causing chromatolysis and further maintaining cell body size and shape in spinal motor neu- enlargement in that particular region of the cell body, rons. In particular, Jacomy et al. [22] have reported that thereby contributing to increased nuclear eccentricity spinal motor neuron cell bodies in transgenic mice lack- away from the axon hillock. It could also act as a force for

Page 9 of 13 (page number not for citation purposes) BMC Neuroscience 2005, 6:19 http://www.biomedcentral.com/1471-2202/6/19

disassembly of the cytoplasmic skeleton on the opposite that is most clearly seen by transmission electron micros- side of the nucleus. copy in resinless thin sections. Nissl bodies are associated with the isolated cytoskeleton, as are lipofuscin granules, Other evidence exists for "damming up" of cytoskeletal which may be caged within the filamentous lattice. proteins destined for the axon. We have previously shown that axotomy increases the content of and neuro- Axotomy increases nucleolar, nuclear and cell body vol- filament protein in the injured cell body, but decreases ume, largely by altering the cytoskeleton. The likely cause their content in the axon proximal to the injury site [10]. of the injury-induced enlargement is the addition of new That study also showed that the volume of the axon, protein to the cell body cytoskeleton. Axotomy also which is especially dependent upon neurofilaments increases nuclear eccentricity by altering the cell body [25,26], decreases after axotomy, while the cell body cytoskeleton. We postulate that both increased nuclear volume is increasing. Those observations, together with eccentricity and central chromatolysis in axotomized the data presented here, lead us to propose that axotomy spinal motor neurons are consequences of damming up causes newly synthesized axonal cytoskeleton to accumu- of Nissl body-free axonal cytoskeleton between the late between the nucleus and axon hillock, resulting in an nucleus and axon hillock. increase in nuclear eccentricity away from the axon hillock and chromatolysis between the nucleus and axon hillock. Methods Surgical procedures A requirement of this proposal is that it explain the direc- All ventral root transections were performed on adult tion in which chromatolysis develops. The loss of southern grass frogs (Rana pipiens) of both sexes, pur- basophilic staining after axotomy is termed "central chased from Carolina Biological Supply, Burlington, NC. chromatolysis" because in many instances it first appears The left 9th and 10th ventral roots were transected 6–9 mm near the nucleus and spreads outward towards the axon from the spinal cord via a dorsal laminectomy under gen- hillock [6]. If the axonal cytoskeleton dams up in this eral anesthesia (135 mg Finquel/kg body wt.). Animals region of the injured cell body, why does the loss of were allowed to survive for up to 20 days before isolation basophilia spread centrifugally? It is not known exactly of motor neuron cell bodies. Pre- and postoperative care where assembly of the axonal cytoskeleton begins in and euthanasia by decapitation were conducted under normal or axotomized motor neurons. However, there is institutional review in accordance with AVMA standards. some information about the assembly of the cytoskeleton in non-neuronal cells. In her pulse-chase analysis of the Cell body isolation incorporation of newly synthesized proteins into the Figure 8 illustrates the procedure used to isolate motor cytoskeletal framework of non-neuronal cells, Fulton [27] neuron cell bodies from unfixed frog lumbar spinal cord found autoradiographic evidence that new cytoskeleton is [10]. Small pieces of R. pipiens lumbar spinal cord were first assembled near the nucleus, and then moves towards first cryoprotected with 70% ethylene glycol and frozen at the periphery of the cell body. If the assembly of the -80°C. Cryoprotection improves the quality and yield of axonal cytoskeleton in motor neurons follows the same motor neurons [28] and permits storage of tissues until pattern after axotomy, then chromatolysis could spread needed. After thawing, the tissue was suspended in 0.9 M from the nucleus to the cell body periphery. As the sucrose in 1.7 mM sodium citrate buffer (pH 5.0) movement of newly synthesized axonal cytoskeleton containing 15 mM glucose and expressed through nylon towards the hillock is impeded, chromatolysis would bolting cloth of 202 µm pore size. Nylon cloth with pores appear first near the nucleus, spreading outward towards of 351 µm on a side was used for the larger human motor the hillock with time and displacing and compressing the neurons. The cell bodies were then centrifuged in a dis- basophilic cytoplasmic skeleton. This action could pro- continous sucrose gradient to obtain a crude neuronal duce a densely basophilic rim in the cell body periphery, fraction, from which individual cell bodies were removed as has often been observed [6]. Further accumulation of with a glass pipette, rinsed twice in 1.5 M sucrose and the axonal cytoskeleton could ultimately spread to other pooled with other purified cell bodies. Motor neurons regions of the cell body. were identified by their distinctive size, shape and bio- chemical profiles [29]. In each experiment, 100–400 cells Conclusion were collected from the lumbar spinal cord of three grass We have shown that nucleolar, nuclear and cell body vol- frogs. ume in spinal motor neurons, as well as nuclear position, is maintained by the cytoskeleton. The isolated cytoskele- Cytoskeleton isolation ton, whose protein composition has not been deter- Cytoskeletons from isolated motor neuron cell bodies mined, contains slightly less than one-half of the total cell were obtained by serial extraction of soluble proteins, body protein. The cytoskeleton is a filamentous lattice using a modification of the method of He et al. [12]. The

Page 10 of 13 (page number not for citation purposes) BMC Neuroscience 2005, 6:19 http://www.biomedcentral.com/1471-2202/6/19

Mince spinal tissue, 1 mm3

7 parts ethylene glycol, 3 parts buffer medium -70°C, minimum of 90 min.

Remove ethylene glycol, express through nylon mesh 202µm2,inbufferedmedium Micropipette

9400 g, 40 min.

0.9M

1.5M

2.1M

IsolationFigure 8 procedure for spinal motor neuron cell bodies Isolation procedure for spinal motor neuron cell bodies. Flow chart of the procedure used to isolate individual spinal motor neuron cell bodies from the frog [10]. Nylon mesh with pore widths of 351 µm2 was substituted for the isolation of human motor neurons.

buffer solution used in each step of the extraction proce- Morphometry dure contained 50 mM HEPES (pH 7.4), 250 mM sucrose, The volume of isolated motor neuron cell bodies was 250 mM NaCl, 10 mM MgCl2, 1.2 mM phenylmeth- quantified using images from a Leitz Diavert microscope anesulfonyl fluoride, and 1 mM EGTA. The sample was equipped with a Fairchild CCD 3000 camera and exposed to 100 µl of ice-cold 0.5% Triton X-100 in analyzed with a Technology Resources Imagemaster 1000. extraction buffer, 35 µl of DNAase (100 µg/ml) treatment Cell body volume was calculated as the product of the for 30 minutes at 24°C., followed by two rinses in 100 µl mean of three measurements of cell body area and the of chilled buffer alone, and extraction with 100 µl of mean of the height at the center and opposite ends of the chilled 0.25 M (NH4)2SO4, and then 2 M NaCl. After each cell body. The cylindroidal shape of the cell body was step, the cell bodies were centrifuged at 1,000 × g for 3 determined from measurements of the z-axis in 120 cells minutes. suspended in 1.2 M sucrose and viewed between two cov- erslips with DIC optics. At the center of the cell body, the

Page 11 of 13 (page number not for citation purposes) BMC Neuroscience 2005, 6:19 http://www.biomedcentral.com/1471-2202/6/19

z-axis averaged 19.3 ± 6.5 µm and was only 102.3 ± 2.2% through the CO2 critical point after 6 wash cycles in a Balz- and 98.6 ± 4.7% of that value, respectively, at opposite ers apparatus. ends of the cell body. In five separate experiments, the z- axis was 72 ± 7% as long as the average cell body radius. Transmission electron microscopy was performed at 80 The flatness of the cell bodies was not induced by pressure kV with a Zeiss EM 10 CA microscope. Negatives were dig- from the apposed coverslips, since it was also observed in itized and image contrast and brightness were adjusted cells tumbling slowly in a droplet of buffer. Nucleolar and electronically. In other experiments, plastic sections were nuclear volumes were calculated as spheres, using the cut from extracted spinal tissue embedded in Spurr's low mean radius from three measurements of nucleolar or viscosity Epon following fixation in 2.5% glutaraldehyde nuclear area. and 1% osmium tetroxide and were post-stained with 1% uranyl acetate and Reynolds' lead citrate. Nuclear eccentricity was measured by the method of Barr and Hamilton [1], using the formula [AC÷(AB-CD)] × Protein analyses 100%. A-D are points along a line from the center of the After TCA precipitation, using a modification of the cell body (A) through the center of the nucleus (C) to the method of Bensadoun and Weinstein [30], total cell body cell periphery (B). AC is the distance between the centers protein content was measured by the micro method of of the cell body and nucleus (each computed from their Lowry et al. [31], standardized to bovine serum albumin. areas by the software program). AB is the distance A total of 200 to 300 cell bodies was used for each analy- between the center of the cell body and its periphery and sis. Triplicate samples of 1.5 M sucrose from the second CD is the distance between the center of the nucleus and rinse step were taken from regions near the purified cell the nuclear lamina. Nuclear eccentricity of zero means the bodies and used as "blanks" for the protein assay. centers of the nucleus and cell body are identical, while 100% eccentricity refers to nuclei in contact with the cell Radiolabeling of motor neuron proteins with 3H-leucine body perimeter. prior to cell body isolation was carried out on unoperated control and 16-day axotomized frogs, using the method Confocal microscopy of McIlwain and Hoke [32]. Briefly, longitudinal sections Human motor neuron cell bodies were isolated from lum- of lumbar spinal tissue from six unoperated frogs or from bar or cervical spinal tissue obtained at autopsy from an the ipsilateral side of six axotomized frogs were incubated individual without neurological disease. Cell body in a Yellow Springs Instrument oxygen monitoring appa- cytoskeletons prepared by the procedure outlined above ratus (Yellow Springs, OH, USA) at 17°C for 16 h in frog were stained for 3 min. with 0.0032% methylene blue and Ringer solution containing 1 mCi/ml of 3H-L-leucine. examined with a Leica TCS-NT confocal microscope. After rinsing and cryoprotecting the radiolabeled tissue, 40–200 cell bodies were isolated as described above and, Electron microscopy in some experiments, their cytoskeletons were obtained Intact, resinless thin sections of cytoskeletal preparations before their radioactivity was measured by liquid scintilla- from isolated motor neuron cell bodies are technically dif- tion counting. ficult to obtain. Thus, lumbar spinal tissue from adult frogs was examined after extraction by the method used Authors' contributions for isolated cell bodies with slight modifications. Seg- DLM designed and supervised the study, participated in ments of longitudinally hemisected lumbar spinal tissue parts of the microscopy, and prepared the manuscript. 1–2 mm long were exposed sequentially to HEPES-buff- VBH, an experienced research technician, performed most ered 0.5% Triton X-100, 0.25 M (NH4)2SO4 and 2 M NaCl of the experiments. and fixed in 2.5% glutaraldehyde. To minimize swelling, the segments were kept on ice and DNAase treat- Acknowledgements ment was omitted. We thank Anahid Kavookjian, Hal Mekeel, Vicky Madden and Bob Bagnell for their assistance with aspects of the electron microscopy and Neil Kra- Embedment was carried out at 72°C in increasing concen- marcy and Michael Chua for their assistance with the confocal microscopy. trations of diethyleneglycol distearate (DGD; Electron This study was supported by grants to DLM from the UNC University Research Council and the UNC Medical Alumni Endowment Fund and by Microscopy Sciences, Fort Washington, PA, USA) from contributions to the Jim O'Shea Fund for ALS Research. which 90% of the acetone-soluble contaminants had been removed. Thin sections were then cut on the DGD- References embedded tissue, placed on Formvar-coated, carbon-sta- 1. Barr M, Hamilton JD: A quantitative study of certain morpho- bilized grids and the DGD removed with dry n-butanol logical changes in spinal motor neurons during axon reaction. J Comp Neurol 1948, 89:93-121. for 1 hour × 3 at 40°C. The resinless sections were then 2. Brattgård S-O, Edström J-E, Hyden H: The chemical changes in transferred stepwise into absolute ethanol and dried regenerating neurons. J Neurochem 1957, 1:316-323.

Page 12 of 13 (page number not for citation purposes) BMC Neuroscience 2005, 6:19 http://www.biomedcentral.com/1471-2202/6/19

3. Edström J-E: Ribonucleic acid changes in the motoneurons of 31. Lowry OH, Rosebrough NJ, Farr AL, Randall RJ: Protein measure- the frog during axon regeneration. J Neurochem 1959, 5:43-49. ment with the Folin phenol reagent. J Biol Chem 1951, 4. Burgess SB, McIlwain DL: Synchronous, intersegmental 193:265-275. responses in motoneurons to lumbar ventral root injury. 32. McIlwain DL, Hoke VB: Radiolabeling motoneuron proteins in NeuroReport 1994, 5:2641-2644. the isolated frog spinal cord preparation. J Neurosci Meth 1994, 5. Chen L, Lund PK, Burgess SB, Rudisch BE, McIlwain DL: Growth 52:197-202. hormone, insulin-like growth factor I, and motoneuron size. J Neurobiol 1997, 32:202-212. 6. Lieberman AR: A review of the principal features of perikaryal responses to axon injury. Int Rev Neurobiol 1971, 14:49-124. 7. Bodian D, Mellors RC: The regenerative cycle of motoneurons, with special reference to phosphatase activity. J Exp Med 1945, 81:469-498. 8. Cammermeyer J: Peripheral chromatolysis after transection of mouse facial nerve. Acta Neuropathol (Berl) 1963, 2:213-230. 9. McHanwell S, Biscoe TJ: The sizes of motoneurons supplying hindlimb muscles in the mouse. Proc R Soc Lond B Biol Sci 1981, 213:201-216. 10. Sinicropi DV, McIlwain DL: Changes in the amounts of cytoskel- etal proteins within the perikarya and axons of regenerating frog motoneurons. J Cell Biol 1983, 96:240-247. 11. Hester SP, Capps-Covey P, McIlwain DL: The plasma membrane of bulk-isolated mature spinal neurons. Brain Res 1978, 159:41-54. 12. He D, Nickerson JA, Penman S: Core filaments of the nuclear matrix. J Cell Biol 1990, 110:569-580. 13. Fulton AB, Wan KM, Penman S: The spatial distribution of polyr- ibosomes in 3T3 cells and the associated assembly of pro- teins into the skeletal framework. Cell 1980, 20:849-857. 14. Penman S: Rethinking cell structure. Proc Natl Acad Sci USA 1995, 92:5251-5257. 15. Jones JCR, Goldman AE, Steinert PM, Yuspa SH, Goldman RD: Dynamic aspects of the supramolecular organization of intermediate filament networks in cultured epidermal cells. Cell Motil 1982, 2:197-213. 16. Lenk RL, Ransom L, Kaufmann Y, Penman S: A cytoskeletal struc- ture with associated polyribosomes obtained from HeLa cells. Cell 1977, 10:67-78. 17. Fey EG, Ornelles DA, Penman S: Association of RNA with the cytoskeleton and the nuclear matrix. J Cell Sci Suppl 1986, 5:99-119. 18. You W, Abe S, Davies E: Cosedimentation of pea root polys- omes with the cytoskeleton. Cell Biol Int Rep 1992, 16:663-73. 19. Meier J, Couillard-Despres S, Jacomy H, Gravel C, Julien JP: Extra neurofilament NF-L subunits rescue motor neuron disease caused by overexpression of the human NF-H gene in mice. J Neuropathol Exp Neurol 1999, 58:1099-1110. 20. Helfand BT, Mendez MG, Pugh J, Deisert C, Goldman RD: A role for intermediate filaments in determining and maintaining the shape of nerve cells. Mol Biol Cell 2003, 14:5069-5081. 21. Troy CM, Muma NA, Greene LA, Price DL, Shelanski ML: Regula- tion of peripherin and neurofilament expression in regener- ating rat motor neurons. Brain Res 1990, 529:232-238. 22. Jacomy H, Shu Q, Cauillard-Sespres S, Beaulieu J-M, Julien J-P: Dis- ruption of type IV intermediate filament network in mice lacking the neurofilament medium and heavy subunits. J Neurochem 1999, 73:972-984. 23. Larivière RC, Nguyen MD, Ribeiro-da-Silva A, Julien J-P: Reduced number of unmyelinated sensory axons in peripherin null mice. J Neurochem 2002, 81:525-532. 24. Peters A, Palay SL, Webster deF H: The Fine Structure of the Nervous Publish with BioMed Central and every System Philadelphia: Saunders; 1976. 25. Friede R, Sarnorajski T: Axon caliber related to neurofilaments scientist can read your work free of charge and in sciatic nerve. Anat Rec 1970, 167:379-388. "BioMed Central will be the most significant development for 26. Hoffman PN, Cleveland DW, Griffin JW, Landes PW, Cowan NJ, disseminating the results of biomedical research in our lifetime." Price DL: Neurofilament gene expression. A major determi- nant of axonal caliber. Proc Natl Acad Sci USA 1987, 84:3472-3476. Sir Paul Nurse, Cancer Research UK 27. Fulton AB: Assembly associated with the cytomatrix. J Cell Biol Your research papers will be: 1984, 99:209s-211s. 28. Sinicropi DV, Hoke VB, McIlwain DL: Isolation of motoneuron available free of charge to the entire biomedical community cell bodies from spinal cord stored at -70°C in ethylene peer reviewed and published immediately upon acceptance glycol. Anal Biochem 1989, 180:286-290. 29. Sato S, Burgess SB, McIlwain DL: Transcription and motoneuron cited in PubMed and archived on PubMed Central size. J Neurochem 1994, 63:1609-1615. yours — you keep the copyright 30. Bensadoun A, Weinstein D: Assay of proteins in the presence of interfering materials. Anal Biochem 1976, 70:241-250. Submit your manuscript here: BioMedcentral http://www.biomedcentral.com/info/publishing_adv.asp

Page 13 of 13 (page number not for citation purposes)